Acute brain injury is complicated by mass effect, a major cause of secondary injury by critically reducing cerebral perfusion pressure or producing tissue shifts which can lead to herniation. While the etiology and mechanisms responsible for this mass effect may differ considerably depending on the clinical condition, treatment options are limited. Traditionally osmotic agents have been the mainstay of medical treatment employing mannitol and recently hypertonic saline solutions. Despite numerous investigations and mannitors ubiquitous use, considerable controversy persists regarding its 1) mechanism of action (whether it reduces cerebral blood volume [CBV] or brain water), 2) differential effect in normal and abnormal brain, 3) effect on cerebral blood flow (CBF) and oxygen metabolism (CMRO2), 4) loss of efficacy over time, and 5) uptake into injured brain, which could lead to "rebound edema." In part because of some of these concerns, hypertonic saline (HS) has been introduced as an alternative agent; however, its physiologic effect and clinical efficacy compared to mannitol is unknown. Optimal use of these osmotic agents to treat brain swelling requires a detailed knowledge of their physiological effects in human subjects under the clinical conditions in which they are used. To this end, we will carry out the following Specific Aims: 1) measure the effect of a single bolus of mannitol or HS (23.4%) on regional CBF, CMRO2, CBV in patients with traumatic brain injury (TBI) and large hemispheric infarctions (LHI). We will measure CBF, CBV and CMRO2 using PET before and after a bolus of equi-osmolar (5.49 mOsm/kg ) doses of mannitol (1.0 gm/kg ) or HS (0.686 ml/kg), in 10 patients each with LHI and midline shift and 10 patients each with severe TBI; 2) measure the effect of a single bolus of mannitol or HS (23.4%) on regional brain volume in patients with LHI. In 20 patients we will measure regional change in brain volume using MRI before and after the same doses of mannitol and HS used in Aim 1. TBI patients will not be studied due to the MRI incompatibility of ICP monitors; 3) measure the extent that mannitol crosses the blood-brain-barrier in patients with TBI and LHI receiving osmotic therapy. In 15 patients with each entity we will measure blood: brain permeability of mannitol using PET and carbon-11 labeled-mannitol.